1. Field of the Invention
The present invention relates generally to semiconductor wafer cleaning and, more particularly, a chuck assembly and a wafer backside plate to be used in semiconductor substrate spin, rinse, and dry (SRD) modules.
2. Description of the Related Art
Particulate contaminants generally consist of tiny bits of distinctly One commonly used wafer preparation operation used at various stages of substrate preparation is a spin, rinse, and dry (SRD) module. Conventionally, the wafer is spin rinsed by spraying deionized water onto the top and backside of the wafer, as the wafer is spun around at high speed. The spin rinse operations are typically performed in a bowl rigidly mounted on an SRD housing designed to receive a spindle coupled to a motor. As the motor rotates, so do the spindle, a chuck mounted on the spindle, and the wafer. Customarily, the chuck supports the edges of the wafer by utilizing four spindle fingers coupled to the chuck. The spindle fingers are designed to move upwardly out of the bowl such that they extends outside the bowl housing. Thus, customarily, the wafer is delivered to the spindle fingers while they are extended out side of the bowl at a level above wafer processing plane. Once the wafer is delivered to the spindle fingers, the chuck having the spindle fingers and wafer attached thereto moves back down and into the bowl so as to place the wafer at the level of wafer processing plane.
Typically, fluid (e.g., DI water) is supplied to a spigot and thus onto the surface of the wafer, as the wafer is spun at high revolutions per minute (RPMs). When the surface of the wafer is sprayed with fluid, the supplying of fluid is stopped by turning off the spigot, and the wafer is dried as the wafer continues to spin at high RPMs. Once the wafer is dried, the processed wafer is unloaded by moving the chuck and spindle fingers holding the wafer upwardly out of the bowl until the wafer is extended above the wafer process plane for a second time. At this time, an end effector can reach in and remove the wafer from the SRD module.
Numerous shortcomings can be associated with chuck assemblies of conventional SRD modules. Primarily, the typical SRD module requires a complex chuck design. For instance, the chuck is commonly required to move up and down. The chuck moves up to receive the wafer, moves down to process the wafer and then up again to remove the wafer from the SRD bowl. In view of this continual movement activity, it is imperative that the chuck remains properly calibrated so that it comes to rest at the exact process level. In situations where the chuck is not properly aligned, failure to properly receive and deliver the wafer mandates the realignment of the chuck. The process of realigning of the chuck is very time consuming and labor intensive, and it requires that the SRD module be taken off-line for an extended period of time, thus reducing throughput.
Another shortcoming of conventional chucks is the unnecessary movements required in loading and unloading of the wafer to and from the fixed spindle fingers. Predominantly, in conventional SRD modules, the loading of the wafer onto the fixed spindle fingers involves four stages. That is, to receive a wafer, initially, the chuck is moved upwardly and out of the bowl, such that the chuck is positioned above the wafer process plane. As a result of having fixed spindle fingers, to deliver the unprocessed wafer to the edges of the spindle fingers, at the outset, the end effector having the wafer is moved horizontally over the bowl at a level that is above the horizontal plane of the spindle fingers (which are already in the up position). Thereafter, the end effector must move downwardly (while over the bowl) until the wafer reaches the level of the spindle fingers. At this point, the spindle fingers can engage the wafer. Once the spindle fingers have engaged the wafer, the end effector relinquishes the wafer and thus physically delivering the unprocessed wafer to the spindle fingers. Finally, to pull out, the end effector is required to move slightly down and away from the wafer under surface before moving horizontally away from over the bowl. Each of the up and down movements of the end effector are performed using the “Z” speed of the end effector, which in fact is a significantly low speed. As such, the performing of a spin, rinse, and dry operation on each wafer requires a significant amount of time simply to load and unload the wafer, thus increasing the SRD cycle per wafer. As can be appreciated, this reduces the overall throughput of the SRD module.
Yet another shortcoming associated with conventional chucks of SRD modules is the creation of air turbulence above the wafer surface. That is, as the chuck and thus the wafer spin in the bowl, the spinning action of the chuck and the wafer transfer energy to air flowing over the topside of the wafer. This transferred energy causes the airflow above the topside of the wafer to become turbulent and thus creates recirculating air (i.e., eddies). The amount of energy transferred to the air flowing to the topside of the wafer depends on the diameter and the rotational speed of the wafer. In general, the greater the amount of energy transferred to the air, the higher the eddies extend above the topside and the farther the eddies extend below the backside of the wafer. The presence of eddies below the wafer is undesirable because particles or DI water droplets removed from the wafer can circulate in the eddies and can be re-deposited on the backside of the wafer, thereby causing wafer recontamination.
Further challenges faced in the use of conventional chucks are the limitations associated with the chuck geometry. Mainly, the relatively large size and associated weight of conventional flat chucks necessitate the use of significantly higher amounts of energy to operate the SRD module. Additionally, the large size of the chuck further requires the use of larger shafts as well as spindles. Collectively, these limitations mandate the use of a larger and more powerful motor, thus increasing the cost of the SRD modules as well as the associated operating cost.
Yet another challenge faced in the use of chucks in SRD modules is having chemically incompatible components present within the modules. In a typical SRD module, most components are constructed from several different materials. For instance, the chuck is usually constructed from Aluminum, while the bowl is made out of Polyurethane, and the spigot is made out of stainless steal. As a result, particles or contaminants from chemically incompatible components may enter into chemical reaction with the fluids introduced into the SRD module, thus further recontaminating the SRD module. This recontamination can further be exacerbated by the aluminum chuck having to continuously move up and down (e.g., to load and unload each new wafer) within the bowl. That is, as the chuck moves up and down within the bowl, some of its coating may flake off of the chuck, thus generating particulates and contaminants inside the bowl and the SRD module. In some cases, these contaminants may react with the residual chemicals (e.g., HF, NH3OH, etc.) present in the SRD module from previous brush scrubbing operations. It is believed that these chemical reactions between the residual chemicals and the generated particulates and contaminants of the chuck may cause the recontamination of the wafer as well as the SRD module.
In view of the foregoing, a need therefore exists in the art for an apparatus that controls and reduces the airflow to a backside of a substrate during a spin, rinse, and dry operations. Additionally, there is a need for a chemically compatible chuck assembly that improves the spin, rinse, and dry operations performed on the surfaces of substrates while reducing the risk of wafer recontamination.